solar versus fission surface power for mars · solar power masses increase as landing sites are...
TRANSCRIPT
American Institute of Aeronautics and Astronautics
1
Solar Versus Fission Surface Power for Mars
Michelle A. Rucker1
National Aeronautics and Space Administration, Lyndon B. Johnson Space Center, Houston, Texas, 77058, U.S.A.
Steve Oleson2, Pat George3, Geoffrey A. Landis4, James Fincannon5, Amee Bogner6, Robert E. Jones7, Elizabeth
Turnbull8, and Jeremiah McNatt9
National Aeronautics and Space Administration, John H. Glenn Research Center, Cleveland, Ohio 44135, U.S.A.
Michael C. Martini10 and John Z. Gyekenyesi11,
Zin Technologies, Middleburg Heights, Ohio, 44130, U.S.A.
Anthony J. Colozza12, Paul C. Schmitz13, and Thomas W. Packard14,
Vantage Partners, LLC, Brook Park, Ohio, 44142, U.S.A.
A multi-discipline team of experts from the National Aeronautics and Space
Administration (NASA) developed Mars surface power system point design solutions for two
conceptual missions to Mars using In-situ resource utilization (ISRU). The primary goal of
this study was to compare the relative merits of solar- versus fission-powered versions of
each surface mission. First, the team compared three different solar-power options against a
fission power system concept for a sub-scale, uncrewed demonstration mission. This
“pathfinder” design utilized a 4.5 meter diameter lander. Its primary mission would be to
demonstrate Mars entry, descent, and landing techniques. Once on the Martian surface, the
lander’s ISRU payload would demonstrate liquid oxygen propellant production from
atmospheric resources. For the purpose of this exercise, location was assumed to be at the
Martian equator. The three solar concepts considered included a system that only operated
during daylight hours (at roughly half the daily propellant production rate of a round-the-
clock fission design), a battery-augmented system that operated through the night (matching
the fission concept’s propellant production rate), and a system that operated only during
daylight, but at a higher rate (again, matching the fission concept’s propellant production
rate). Including 30% mass growth allowance, total payload masses for the three solar
concepts ranged from 1,128 to 2,425 kg, versus the 2,751 kg fission power scheme. However,
solar power masses increase as landing sites are selected further from the equator, making
landing site selection a key driver in the final power system decision. The team also noted
that detailed reliability analysis should be performed on daytime-only solar power schemes
to assess potential issues with frequent ISRU system on/off cycling.
1 Engineer, Exploration Integration and Science Directorate, Mail Stop XM. 2 COMPASS Team Lead, Mission Architecture and Analysis Branch (LSM), Mail Stop 162-2. 3 Engineer, Exploration Systems Project Office (MX), Mail Stop 162-7. 4 Research Engineer, Photovoltaic and Electrochemical Systems Branch (LEX), Mail Stop 302-1. AIAA Associate
Fellow. 5 Electrical Power Systems Engineer, Power Architecture and Analysis Branch (LEP), Mail Stop 302-1. 6 Aerospace Technologist, Avionic Systems (LEA), Mail Stop 309-2. 7 Communications Engineer, Communications and Intelligent Systems Division (LC00), Mail Stop 54-4. 8 Operations Research Analyst, Cost and Economic Analysis Office (BC), Mail Stop 77-7. 9 Aerospace Technologist, Photovoltaic and Electrochemical Systems Branch (LEX), Mail Stop 302-1. 10 Aerospace Engineer, Mission Architecture and Analysis Branch (LSM), Mail Stop 162-2. 11 Mechanical Engineer, Mission Architecture and Analysis Branch (LSM), Mail Stop VPL-3. 12 Aerospace Engineer, Materials and Structures Division (LSM), Mail Stop 309-1. 13 Engineer, Thermal Energy Conversion Branch (LET), Mail Stop 301-2. 14 Aerospace Engineer, Mission Architecture and Analysis Branch (LSM), Mail Stop 162-2.
https://ntrs.nasa.gov/search.jsp?R=20160010550 2020-04-19T16:51:21+00:00Z
American Institute of Aeronautics and Astronautics
2
Next, the team developed a solar-powered point design solution for a conceptual four-
crew, 500-day surface mission consisting of up to four landers per crewed expedition
mission. Unlike the demonstration mission, a lengthy power outage due to the global dust
storms that are known to occur on Mars could pose a safety hazard to a crewed mission. An
earlier power trade study performed by NASA in 2007 concluded that fission power was
more reliable—with a much lower mass penalty—than solar power for this application.
However, recent advances in solar cell and energy storage technologies and changes in
operational assumptions prompted NASA to revisit the analysis. For the purpose of this
exercise a landing site at Jezero Crater, located at 18o north latitude, was assumed. A fission
power system consisting of four 10-kW fission reactors was compared to a distributed
network of Orion-derived solar arrays mounted on every lander, plus a complement of
Lithium Sulfur batteries. The team found that a solar power generation and storage system
mass of about 11,713 kg on the first crewed expedition would provide the 23 kilowatts (kW)
needed to survive a dust storm lasting up to 120-days at average optical depth of 5; note that
this may be insufficient to complete return vehicle propellant production. The arrival of
6,102 kg more solar power generation equipment on the second crewed expedition would
match fission power’s ability to operate at full power, even during a lengthy dust storm, but
at nearly twice the cumulative landed mass of the 9,154 kg fission system.
Relative merits—and risks--of the two approaches were also evaluated. To assess latitude
sensitivity, the team performed a brief assessment of a second notional landing site,
Columbus Crater, located at 30o south. Because global dust storms appear in the Southern
hemisphere during the summer, rather than the winter, when the days are shorter, solar
array mass did not increase at Columbus Crater—but more battery mass would be required
for the longer nights, and additional ISRU strings would be needed for optimal
summer/winter cycling. Based on this work, it is clear that the choice between fission versus
solar power for Mars surface systems is an important decision, with potential ramifications
to crewed landing site selection.
Nomenclature
cg = center of gravity
K = Kelvin
kg = kilograms
kg/hr = kilograms per hour
km = kilometer
kWe = Kilowatt electric
m = meters
m2 = square meters
mR/hr = millirem per hour (radiation exposure rate)
rad = 0.01 Joules per kilogram (absorbed radiation dose)
sol = solar day on Mars (24 hours 39.5 minutes)
= atmospheric optical depth
V = Volt
W-hr/kg = Watt-hours per kilogram
I. Introduction and Background
ARLY Mars mission concepts developed by the National Aeronautics and Space Administration (NASA)
estimated that 40 kilowatts electric (kWe) power would be needed to support six crew living on Mars for about
500 days.1 To minimize mass, the crew’s return vehicle would land on Mars with empty oxygen propellant tanks
and a manufacturing plant would produce Liquid Oxygen (LOX) propellant from the Martian atmosphere. Analysts
traded solar versus nuclear power, and settled on a fission system, primarily due to the possibility of global dust
storms that threatened solar power’s reliability. Initial studies focused on a monolithic, 40-kWe Fission Surface
Power System (FSPS) concept2 that was being considered by the Constellation Program for lunar applications. The
FSPS would be deployed about one kilometer (km) from the eventual habitat location and used to power an In Situ
Resource Utilization (ISRU) propellant plant during the cargo mission phase; once the return vehicle’s propellant
E
American Institute of Aeronautics and Astronautics
3
tanks were full, the crew would arrive and switch the FSPS over to habitat and science support during the crewed
mission phase. Later work3 demonstrated advantages—including mass savings—of breaking the 40-kWe FSPS into
multiple, smaller power sources by leveraging a joint NASA-Department of Energy (DoE)/Los Alamos National
Laboratory (LANL) research effort called the Kilopower fission system concept.4 The lower mass Kilopower units
were assumed to make fission power even more compelling than solar power for this application.
Before initiating the expensive process to develop and flight certify a fission system, NASA’s Evolvable Mars
Campaign (EMC) managers thought it prudent to refresh the solar versus fission power trade study to include recent
technical advances in solar panel efficiency and energy storage density, as well as updates to mission operational
concepts. This task was assigned to the Collaborative Modelling for Parametric Assessment of Space Systems
(COMPASS) team at NASA’s John H. Glenn Research Center.
II. Study Approach
The study was broken into two individual tasks. First, the COMPASS team traded solar power against fission
power for an uncrewed, subscale ISRU demonstration mission. The primary objective of this mission would be to
test Mars entry, descent, and landing technologies but with a secondary goal of demonstrating that cryogenic
propellants can be manufactured from Martian resources within the prescribed mission timeline. The oxygen
production rate for this demonstration was specified to be 1/5 of the estimated production need for a human mission.
Next, the COMPASS team revisited the power trade study for a crewed mission, using EMC mission assumptions5
and state of the art technologies.
In the first part of the study, power outages due to dust storms were factored in, but spacecraft power could be
reduced without risk; in the second part of the study, the system had to maintain minimum crew keep-alive power
regardless of dust storm severity or duration.
III. ISRU Demonstrator Analysis
A. ISRU Demonstrator Task Approach
The first challenge the study team encountered was how to perform an “apples to apples” comparison between
the solar and fission powered concepts, given that the solar option would be sensitive to day/night cycling and dust
storms whereas the fission option would operate around the clock without weather-related disruption. The objective
of the study was to compare fission and solar power systems for the same rate of production so the team decided to
evaluate three different solar power cases: Option 1A assumed daytime-only solar-powered production, for
approximately 10 hours of production per day, at the 1/5 scale production rate of 0.45 kg/hr; Option 1B assumed the
same production rate, but with around-the-clock production; and Option 1C assumed day-time only production, but
at twice the production rate of Options 1A or 1B. These three solar options were then compared against Option 2, a
1/5 production rate fission system operating around the clock. The team selected five Figures of Merit for option
comparison:
1) Performance, as measured by propellant production rate in kilograms per hour (kg/hr).
2) System mass.
3) System robustness to the Martian environment, particularly dust.
4) System design, development, test, and evaluation cost.
5) System operational life.
B. ISRU Demonstrator Task Assumptions
1. Launch and Landing
The COMPASS team assumed the demonstrator payload would launch on a Delta IV Heavy, sized to fit within a
five meter diameter shroud, with a total payload launched mass of 7,500 kilograms. A seven meter diameter
Adaptable Deployable Entry and Placement Technology (ADEPT) entry system was assumed6. Subtracting
estimated entry, descent and landing systems mass based on recent Evolvable Mars Campaign analysis left about
3,500 kg available for the ISRU demonstration experiment, including power systems. The demonstrator spacecraft
was assumed to manufacture propellant from the carbon-dioxide atmosphere, rather than soil; therefore the mission
could be conducted anywhere on the planet. In lieu of defined mission details, the team assumed the Mars
Exploration Rover (MER) Program’s Opportunity rover landing site at Meridiani, roughly 2o south, as this had the
benefit of Opportunity’s 12 years of actual measured solar array performance data, as well as favorable night
durations and minimal seasonal variations throughout the year.
American Institute of Aeronautics and Astronautics
4
2. Martian Environment
To size thermal systems, the study team used MER data from the Opportunity landing site. Surface temperatures
were assumed to vary from a night time low of 174 K to a daytime high of 308 K; sky temperatures were assumed to
swing from 139 K to 213 K.
Dust storms on Mars include both local dust storms and occasional global storms. While the local storms last at
most a few days and have only a modest increase in optical depth, a global storm can last for months, with a
significant increase in optical depth7. Such global storms seem to occur, on average, about every three Martian
years, and are more prevalent during the southern hemisphere spring and summer. For this study, a dust storm of
120 days in duration was assumed, also based on data collected from various Mars observation programs8. For the
purpose of this study, optical depth (, the logarithm of
the ratio of incident to transmitted radiant power on a
vertical path through the atmosphere) was assumed to
vary from 1.0 under clear skies to a worst-case value of
5.0 during a dust storm, based on solar-energy
measurements taken by the Spirit and Opportunity
rovers. Maximum wind speeds up to 20 meters per
second were assumed. Figure 19 illustrates the visible
difference of various optical depths, as viewed by the
Opportunity rover. While it may appear that almost no
sunlight was reaching the rover at the peak of the storm,
only direct sunlight was blocked. The dust caused light
to be scattered, such that at the worst of the storm the
diffuse light was approximately 30-40% the amount of
direct light on a clear day. It was assumed that, on
average, the arrays would see 12 hours of sunlight per
sol.
3. ISRU System
Preliminary analysis indicated that an ISRU
system sized to produce 0.45 kg/hr could be readily
developed, effectively providing a 1/5 scale
demonstrator for the 2.2 kg/hr full-scale crewed
propellant production needs. ISRU system design and
optimization was out of scope for this exercise, so the
study team assumed a representative ISRU system10,11
(Figure 2), consisting of a carbon dioxide collection
and freezing subsystem, solid oxide electrolysis
processor and recirculation, oxygen liquefaction and
storage, power storage and conditioning equipment,
and a thermal management subsystem.
For the purpose of this exercise, it was assumed
the payload must demonstrate liquefaction and storage of oxygen at cryogenic temperatures. A storage system
capable of holding 1,500 kg of LOX was assumed for all demonstrator options. Spacecraft systems were assumed to
be single-fault tolerant where appropriate.
4. Solar Arrays
For all solar-powered demonstrator concepts, the study team assumed a derivative of the 120 V Orbital ATK
UltraFlex™ arrays or equivalent, populated with Inverted Metamorphic Multi-junction (IMM) solar cells of 33%
conversion efficiency (measured at Earth distance solar flux, 28°C, beginning of life). Array and gimbal concepts
originally developed for the Orion spacecraft’s 2.7 standard gravity (g) Trans Lunar Injection structural loads were
assumed to be sufficient for Martian gravity and surface wind loads. Arrays were sized based on average solar
distance of 1.52 astronomical units, while acknowledging that power could vary as much as +/- 20% depending on
actual distance from the Sun during the Martian year. Arrays were sized to meet the daytime ISRU production and
housekeeping power requirements plus the power needed to recharge energy storage for nighttime housekeeping
power. For worst-case dust storms and resulting incident energy reduction to 35% of the baseline design, solar
arrays were sized to provide more than adequate power to meet minimum loads. A permanent dust power loss
Figure 2. Representative ISRU System.
Figure 1. Dust storm time lapse photo taken by
Opportunity.
American Institute of Aeronautics and Astronautics
5
factor of 0.95 was assumed, given that the two-axis gimbal articulation allowed arrays to periodically rotate to
remove most dust accumulation.
5. Kilopower System
The study team assumed a full-scale, 10-kWe Kilopower system for the fission-powered demonstrator case.
Although this turned out to be slightly oversized for the demonstration mission, it had the advantage of providing
test data for an eventual crewed mission, without the time or expense of developing a single-use sub-scale unit. A
stowed 10-kWe Kilopower system measures approximately 4 m
tall by 1.5 m diameter; with the 20 square meter (m2) radiator
panel deployed, the diameter expands to approximately 5 m, as
shown in Figure 3. However, for the purpose of this exercise, a
simplified unit with a fixed, conical radiator was used. System
mass, including 15% growth allowance, was estimated at 1,754 kg.
Note that nearly a third of the Kilopower’s mass was dedicated to
a shield sized to reduce crew radiation exposure to less than 3
millirem per hour (mR/hr) within 500 meters. As noted, the
Kilopower concept is a joint NASA/DoE technology development,
and was not developed by this study team.
C. ISRU Demonstrator Concept of Operations
The concept of operations for all four options begins the same: launch on a Delta IV Heavy or equivalent and
land at the Opportunity site, 2o south of the Martian equator, in late 2025. The Kilopower option would perform
system checkouts for approximately two weeks and then initiate propellant production; the three solar-powered
options would first deploy solar arrays before performing system checkouts and beginning propellant production.
The solar-powered options would gimbal their solar arrays to
follow the sun, and periodically articulate the arrays +/- 45o (Figure
4) to clear accumulated dust. Note that there is a practical
limitation on solar array diameter, to prevent ground contact, or
contact between multiple arrays, during gimbaling.
In the event of a dust storm, the three solar powered options
would pause propellant production and use stored energy to
maintain accumulated propellant at cryogenic conditions. For all
options, production would continue until 4,400 kg of LOX was
produced (1/5 of the estimated crewed mission need). However,
only 1,500 kg of LOX would be stored and conditioned in the
accumulator; the balance would be vented overboard.
D. ISRU Demonstrator Conceptual Design
1. ISRU production plant
The basic ISRU plant was the same for each of the options studied. As noted above, the study team leveraged
work previously performed by NASA to estimate ISRU production plant masses, with one exception: the cryopump
used for atmospheric carbon dioxide collection was replaced with a scroll pump to reduce both radiator mass and
volume. This change brought the ISRU subsystem mass down by almost 50 kg.
The Command and Data Handling subsystem was comprised of a Maxwell Technologies SCS750® space-rated
computer, an Aitech 710 radiation tolerant communications card, an enclosure and power supply, Ethernet Local
Area Network (LAN) card, solar array gimbal controller, memory storage unit, sensors, instrumentation, and wiring.
The Communications and Tracking subsystem consisted of two external Ultra High Frequency (UHF) band
omnidirectional antennas and an internal UHF L3 proximity link transceiver for communications to an orbiting relay
station. A nominal data rate of 100 kilobits per second (kbps) was assumed, with communications occurring for
about an 8 minute window per each sol.
The thermal control system was dominated by a pumped loop cooling system and radiators to remove heat from
the oxygen liquefaction cryocooler. To maintain internal electronic components within an operating temperature
range of 233 K to 323 K, electronics boxes were protected with Aerogel foam insulation, thermal paint was applied
to exposed surfaces, and radiators were used for cooling internal electronics.
Figure 4. Gimbaled solar arrays.
Figure 3. Kilopower Concept.
American Institute of Aeronautics and Astronautics
6
Finally, a 7075-T73 Aluminum and composite sandwich
primary structure provided the backbone for the conceptual
design (Figure 5). The aluminum and carbon fiber-reinforced
cyanate ester matrix structure was designed to carry axial
launch loads as high as 6.5 g axial and 0.5 g lateral, lateral
launch loads up to 2.5 g axial and 2.0 g lateral, and up to 12 g
axial loads during landing parachute deployment. The
estimated structures mass could be optimized with a higher
fidelity stress analysis and the use of more advanced
composite materials, orthogrid, or isogid panels.
2. Solar Option 1A: Daylight-Only, 1/5 Rate Production
The study team estimated that the spacecraft required
about 0.22 kW power for post-landing system checkout and
array deployment, but the load jumped to 6.45 kW once propellant production began. Night time standby mode was
0.69 kW. The study team calculated that four each 5.6 m diameter UltraFlex™ arrays or equivalent, estimated at 30
kg each, would provide the power needed to produce 4,400 kg of LOX for an ISRU system operating 10 hours per
sol (assuming 2 hours per sol of sunlight lost to system warm-up). Lithium-ion batteries would provide standby
power for communications and thermal control while the ISRU system cycled off through the night. Including the
possibility of pausing production during a dust storm of up to 120 sol duration, propellant production would require
about 1,098 sols. Based on a 6 m diameter landed footprint, the maximum center of gravity (cg) height was
approximately 2.61 m for this concept. Overall height was 4.61 m.
A breakdown of the 1,128 kg total payload mass, including growth allowances12, is shown in Table 1.
In addition to the solar arrays, the power subsystem included 60% depth of discharge, Panasonic cell type
Lithium-ion batteries, with an energy storage density of 165 Watt-hours per kilogram (W-hr/kg). A power converter
derived from International Space Station heritage hardware would convert the 120 V power provided by the solar
arrays to 28 V for other spacecraft loads. A 120 V solar array regulator based on a commercial proprietary design
was sized to handle the peak power loads expected at the Mars surface and optimize the power obtained from the
solar array. 28 V power distribution cards provide switching to spacecraft loads and 28 V battery charge-discharge
cards provides battery regulation and protection. Estimated masses for a 28 V battery controller and power
distribution system were based on commercial TERMA hardware. Wire harness mass was sized at 13% of power
system mass.
The final spacecraft landed footprint was approximately 6.0 m. With all four arrays deployed the maximum
diameter was 15.6 m and 4.81 m tall (with arrays not gimbaled).
3. Solar Option 1B: Around-The-Clock, 1/5 Rate Production
Switching to around-the-clock ISRU production cut the mission time in half, but required larger solar arrays and
more energy storage, causing the electrical power subsystem alone to jump more than 1,200 kg in mass as compared
to the 1A daylight-only option. Each of the four 7.5 m diameter solar arrays needed for this option was more than
twice the mass of the 5.6 m diameter arrays specified for Option 1A, but the bigger impact was battery mass to
continue ISRU production through the night: over 1,100 kg more batteries for Option 1B, as compared to Option
1A. This option's ISRU plant, command and data handling, communications and tracking, thermal control, and
structures subsystems were all identical to the 1A Option but electrical subsystem changes resulted in a 2,425 kg
Table 1. Demonstrator Option 1A manifested equipment list.
Description Basic
Mass (kg)
Growth
(%)
Growth
(kg)
Total Mass (kg)
ISRU Payload 163 18 29 192
Command and Data Handling 45 16 7 52
Communications and Tracking 6 17 1 7
Electrical Power Subsystem 351 30 104 455
Thermal Control (Non-Propellant) 140 18 25 165
Structures and Mechanisms 155 18 28 183
Mars ISRU Demonstrator Sub-Total 860 23 194 1054
System-Level Growth 7 74 74
Mars ISRU Demonstrator 1054 25 268 1128
Figure 5. ISRU Demonstrator Option 1A, shown
with solar arrays stowed for landing.
American Institute of Aeronautics and Astronautics
7
total payload mass, more than double the 1A option. Figure 6
shows how much larger the 1B spacecraft footprint would be
with the larger solar arrays, as compared to the 1A option.
4. Solar Option 1C: Daylight Only, 2/5 Rate Production
The third solar-powered ISRU Demonstrator option
utilized the larger 7.5 m diameter solar arrays to power an
ISRU system twice the capacity of the other two options, but
only during the approximately 10 hours per sol of daylight.
This scheme caused peak power to jump to 12.43 kW, nearly
double that of the 1A option. At 1,531 kg, payload total mass
fell between the 1A and 1B options. The command and data
handling, communications and tracking, and thermal control
subsystems were all identical to the other two solar powered
payload concepts, but this option carried 143 kg more ISRU
mass and 37 kg more structural mass than the other two
options.
5. Fission Option 2: Around-The-Clock, 1/5 Rate Production
With a fixed, conical upper radiator, the fission-powered
option (Figure 7) required no post-landing mechanism
deployment and needed only 106 W keep-live power after
landing. Peak power draw of 6.52 kW is achieved during night-
time operations, with a peak daytime power need of 6.45 kW.
Based on a 6 m diameter landed footprint, the maximum cg
height was the same as the solar options at 2.61 m, but overall
landed height was slightly higher, at 5.14 m.
At 2,751 kg total payload mass (including growth
allowance), this option is considerably larger than the 1A or 1C
solar power options, but comparable to the 1B option. As with
the 1B option, the higher mass is almost entirely due to the
electrical power system's 1,804 kg allocation, which is
dominated by a 1,754 kg, 10-kWe Kilopower unit. The
command and data handling, communications and tracking, and thermal control subsystems were all identical to the
solar powered payload concepts; the thermal control system was actually lower mass than the solar options.
E. ISRU Demonstrator Solar versus Fission Comparison
Table 2 compares key parameters of the four demonstrator mission options.
1. Performance Comparison
The closest “apples to apples” comparison with fission is the 1B solar around-the-clock option, as both are
producing propellant at the same rate (though it could take the solar option longer to do so given the probability of a
dust storm). A mission using the 1A concept would take twice as long as the other missions.
2. Mass Comparison
Options 1B and 2 are roughly comparable in mass for an equatorial landing, though the solar concept mass
would quickly exceed the fission option mass as the landing site shifts further from the equator. All four options fit
comfortably within the 3,500 kg landed payload allocation. The 1A and 1C options could both accommodate two
metric tons of additional payload to the surface. All four options would provide flight experience for the ISRU
system but Option 2 would also provide flight experience for a full-scale Kilopower unit. Unfortunately, this
handicaps the fission option in a mass trade against the solar options because a full-scale Kilopower is about 30%
oversized for this particular mission. What’s more, a baseline Kilopower unit includes sufficient radiation shield
mass to protect crew within one kilometer, which is not necessary for this uncrewed mission.
Figure 7. Fission-powered ISRU demonstrator
concept.
Figure 6. Solar array size comparison between
1A daytime-only (left) and 1B around-the-clock
ISRU (right) production.
American Institute of Aeronautics and Astronautics
8
3. System Robustness
Dust is a major concern for Martian surface operations. Solar array deployment and gimbal mechanisms could be
compromised by fine grit, leading to reduced performance for options 1A, 1B or 1C. Similarly, dust accumulation
on the conical radiator could lead to reduced performance for option 2. The study team included solar array gimbals
to remove dust, based on experience with angled surfaces on NASA’s previous Mars rover programs. ISRU system
tolerance to repeated cycling for the 1A and 1C options is unknown. In all cases, ISRU system tolerance to fine dust
ingestion would have to be understood and potentially mitigated.
4. Cost Comparison
A rough order of magnitude cost comparison was assembled for solar option 1B and fission option 2 (Table 3).
Estimates include only Phase B/C/D13 lander costs, and do not include launch services, National Environmental
Protection Agency and Nuclear Safety Launch Approval (where applicable), mission-level costs, the ISRU payload
and ISRU integration cost, fuel costs, science and science instruments, or technology development up to Technology
Readiness Level (TRL) 6. These estimates also ground rule out technology development costs for the fission power
system and solar arrays to allow for a better comparison, as these may be carried under a separate program.
Estimates assume protoflight development and the effort is contracted to a major aerospace firm. Quantitative risk
analysis was performed using a Monte Carlo simulation driven by input parameter uncertainty and the error statistics
of the cost estimating relationships. Costs presented are mode values (approximately 35th percentile), in constant
United States (U.S.) Fiscal Year 2016 dollars (FY16$).
5. Operational Life Comparison
As noted, a mission using the 1A concept would take twice as long as the other missions, with option 2 having
the shortest mission timeline. Extended life analysis was beyond the scope of this study but, in theory, all concepts
could remain functional for years after the demonstrator mission is complete. For eventual crewed mission use, the
Kilopower unit would be rated for at least 12 Earth years’ service life. Solar arrays could be expected to have
graceful degradation, allowing reduced performance beyond the planned mission with lost strings or even a lost
wing. Battery performance would decrease over time. Long-term performance of electronics and gimbals are
unknown.
Table 2. Demonstrator Options Comparison.
Option
Solar 1A: 1/5
rate Daytime
Only
Solar 1B: 1/5 rate
Around the Clock
Solar 1C: 2/5
Rate Daytime
Only
Fission 2: 1/5 Rate
Around the Clock
Fission Power
Total Payload Mass
(including growth) 1,128 kg 2,425 kg 1,531 kg 2,751 kg
Electrical Subsystem Mass 455 kg 1,733 kg 639 kg 1,804 kg
ISRU Subsystem Mass 192 kg 192 kg 335 kg 192 kg
Power ~8 kW Daylight
~8 kW Continuous
(with 16 kW of
arrays)
~16 kW Daylight ~7 kW Continuous
Solar Arrays 4 each x 5.6 m
diameter 4 each x 7.5 m dia.
4 each x 7.5 m
diameter None
Night Production? No Yes No Yes
LOX Production 4.5 kg/sol 10.8 kg/sol 9.0 kg/sol 10.8 kg/sol
Time to Produce 4,400 kg
LOX, including 120-Day
Dust Storm Outage
1,098 sols 527 sols 609 sols 407 sols
ISRU On/Off Cycles 1,098 <5 609 <5
Radiation Tolerance 100 kilorad (krad) electronics and ISRU
300 krad
electronics, 10
Mega ad (Mrad)
ISRU
American Institute of Aeronautics and Astronautics
9
IV. Crewed Surface Mission Assessment
A. Crewed Mission Task Approach
The study team leveraged demonstrator mission results for the crewed mission task, and applied the EMC
concept of operations. Solar- versus fission-powered concepts were compared at two hypothetical landing sites, at
different distances from the equator. Although all options were sized to produce crew return propellant, that portion
of the mission could more easily tolerate dust storm disruption because it would occur before the crew arrived. The
more critical power need was after crew arrival, when at least 14.8 kW of crew and equipment keep-alive power
were needed to continue the surface mission and maintain return vehicle propellant conditioning.
B. Crewed Mission Task Assumptions
1. Launch and Landing
The study team assumed that crewed mission elements would launch on NASA’s Space Launch System, sized to
fit within a 10 m diameter payload shroud. To understand solar power mass sensitivity with respect to latitude the
study team selected two hypothetical landing sites (indicated by stars in Figure 8): Jezero Crater, located at 18.9o
North, 77.5o East, and Columbus Crater, located 29.5o South, 166.1o West. Jezero Crater was an early candidate
during the Mars Science Laboratory landing site selection process and is among the sites being considered for the
upcoming Mars 2020 mission14,15; Columbus Crater provided a convenient data point at a mid-latitude in the
southern hemisphere.
2. Surface Mission
Per EMC working assumptions, surface infrastructure (including power systems) would be re-used by
subsequent crews at a single landing site. The first three cargo landers would be pre-positioned before the first
expedition crew’s arrival on the fourth lander, after their Mars Ascent Vehicle’s (MAV) ISRU propellant tanks were
confirmed full. Subsequent expeditions would require only two precursor landers—one for cargo and one for their
ascent vehicle—plus a crew lander, making a total of ten landers for the first three crewed expeditions during the
first decade on Mars. Each expedition crew would spend approximately 500 days on the surface. High voltage
Table 3. Solar Option 1B versus Fission Option 2 Flight Hardware Cost Comparison (Millions of Dollars).
Work Breakdown
Structure/Description
Case 1B (FY16$M) Case 2 (FY16$M) Total
Delta
FY16$M
DDT
&E
Flight
Hardware Total
DDT
&E
Flight
Hardware Total
06.1.3 Command &Data
Handling 26.6 9.4 36.0 26.3 9.4 35.7 0.3
06.1.4 Communications and
Tracking 3.5 1.6 5.1 3.5 1.6 5.1 0.0
06.1.5 *Electrical Power
Subsystem 36.2 80.9 117.1 23.6 175.7 199.3 -82.2
06.1.6 Thermal Control
(Non-Propellant) 10.0 6.4 16.4 8.4 5.3 13.7 2.7
06.1.11 Structures and
Mechanisms 7.8 10.2 17.9 8.0 11.1 19.1 -1.2
Subsystem Total 84.1 108.5 192.6 69.8 203.1 272.9 -80.3
Systems Integration
Integration, Assembly,Check-out
/Systems Test Operations
11.3 6.8 18.1 9.7 9.9 19.6 -1.5
Ground Support Hardware 14.2 14.2 11.8 11.8 2.4
System Engineering/Integration 14.5 22.6 37.1 12.5 33.0 45.6 -8.5
Project Management 10.6 13.7 24.4 9.5 20.2 29.8 -5.4
Launch and Orbital Operations
Support
11.8 11.8 9.8 9.8 2.0
Spacecraft Total 146.5 151.7 298.1 123.1 266.3 389.3 -91.2
Fee (10%) 29.8 38.9 -9.1
Prime Total 328.0 428.3 -100.3
*for comparison purposes, solar array and fission power system technology development costs have been omitted.
American Institute of Aeronautics and Astronautics
10
transmission cables laid on the surface would carry power from point of origin to the various landers. Crew would
live primarily in a surface habitat, with frequent excursions in a pressurized rover to explore surrounding areas.
During the crew’s stay, the surface power system would continue to condition cryogenic propellants and maintain
system health and status checks on the MAV.
3. Martian Environment
As with the demonstrator mission, at least one dust storm of 120 days in duration was assumed, but seasonal
effects were also considered. Sites in the northern hemisphere are closer to the Sun in the winter—when dust storms
are more prevalent in the north—so distance from the Sun counterbalances day/night cycle effects, effectively
maintaining an energy balance similar to an equatorial site. Dust storms in the southern hemisphere are more
prevalent in the summer, when days are longer and the site is closer to the Sun, so distance from the Sun does not
counterbalance day/night cycle effects. To take advantage of this, the study team assumed higher ISRU rates during
summer for the solar-powered option at the Columbus Crater site, as this eliminated the need to add more solar
arrays for the short day/far Sun part of year.
For the Jezero Crater site, the study team assumed night periods up to 13.3 hours, daylight periods as short as
11.4 hours, nominal optical depth of 1.0, and dust storm optical depth of 5.0, or 35% of nominal incident energy. For
the Columbus Crater site, the study team assumed night periods as long as 14.5 hours, days as short as 10.2 hours,
nominal optical depth of 0.8 during propellant production and 1.0 during the crewed mission, and dust storm optical
depth of 5.0.
4. Surface Systems
For the purpose of this study, the team assumed that during the cargo phase the atmospheric ISRU system must
produce 22,728 kg of LOX within 14 Earth months (or 420 days), for a minimum production rate of 2.2 kg per hour
(not accounting for outages). Note that this may be a conservative estimate for some mission scenarios. For a
hypothetical launch in May, 2038 of the first lander, the ISRU system could have as long as 725 to 750 Earth days to
manufacture the first crew’s return propellant. Production schedule for subsequent expeditions varies, but some
expedition timelines could be as short as only 430 Earth days. This analysis only considered atmospheric ISRU;
more power would be needed for regolith excavation and processing to produce soil ISRU.
Once crew arrived, LOX would continue to be thermally conditioned in the MAV propellant tanks, along with
the methane transported from Earth. Crew would live in a separate surface habitat, from which most other
Figure 8. Hypothetical crewed mission landing sites on Mars.
American Institute of Aeronautics and Astronautics
11
equipment—such as space suit batteries or robotic rovers—would be recharged. An optional science laboratory
would be available for sample processing, analysis, and return preparation
As shown in Table 3, estimated peak surface power need during the ISRU-production cargo phase for this
assumed complement of equipment is about 26.3 kW, but could increase to 31 kW during the crewed phase. Note
that the peak crewed power load assumes all
systems operating, including an optional science
facility, but with careful phasing the crewed phase
peak power could be lowered to match the cargo
phase. During the cargo phase, peak and keep-alive
power are the same, since propellant must still be
produced. During the crewed phase, life support
and MAV propellant conditioning take priority but
it was assumed that non-essential functions
(laundry or science experiments, for example)
could be suspended to minimize keep-alive power.
Also note that these values are based on conceptual
element designs, and may increase or decrease as
mission requirements are defined.
5. Kilopower System
The study team assumed the same 10-kWe Kilopower system as the demonstrator mission, but assumed the
EMC baseline of four units plus one spare, all arriving on the first lander. A power distribution system connected the
power farm to subsequent landers. Kilopower mass used in this analysis assumes a one kilometer safe separation
distance between the power farm and the crew habitat. Note that Kilopower mass could be reduced by reducing
shield mass, but this would increase separation distance (and therefore distribution cable mass and line losses).
6. Solar Arrays
For the solar powered concepts, the study team took a novel approach by assuming solar arrays on every lander,
rather than concentrated on a single power farm, as was assumed for the fission case. Power distribution cables
made each lander a node in a large power network, with batteries providing overnight power. Note that dust storm
power was provided by excess solar array area, not energy storage. Given enough landers, the accumulated array
area could collect enough energy even in a dust storm to meet peak demand, without having to reduce loads to “keep
alive” power. Although the first few landers would be at higher risk for dust storm disruption, one appealing
advantage was that power generation capability on every lander made most landers self-sufficient (at least in keep-
alive mode) immediately after landing, reducing the time pressure to autonomously connect to the power farm. This
scheme also made higher latitude sites more accessible for solar power, assuming multiple expeditions contribute to
the network.
As with the demonstrator concepts, the study team assumed crewed mission solar panels were a derivative of the
120 V Orbital ATK UltraFlex™ arrays, with two-axis gimbals to remove dust. Solar power generation assumed a
daily average flux and worst day of year conditions, rather than an hour-by-hour or day-by-day simulation. A Roll-
Out Solar Array (ROSA) option was examined, but the design was not sufficiently mature for the large size, wind
loading, and Mars gravity. A ROSA laid flat on the surface would incur mass penalty for additional array area (since
it could not track the sun), and would be at significant risk of performance degradation due to dust accumulation.
Additional structure could raise the arrays up, but with more mass penalty. It simply made more sense to repurpose
the lander structure as a permanent solar array platform, particularly since most landers served no other purpose
once cargo was off-loaded.
Because each lander was assumed to carry at least one sol worth of internal batteries for on-board keep-alive
power, it was assumed that these batteries could be recharged via the solar arrays during the day to provide night-
time keep-alive power at each lander. Given the longer technology development window for a crewed mission
versus the demonstrator mission, the study team assumed higher energy density 120 V, 270 W-hr/kg Lithium Sulfur
batteries for energy storage. The entire complement of batteries required for full night time power was assumed to
be manifested on the first lander.
7. Power Management and Distribution
For the purpose of this study, it was assumed that the Power Management and Distribution (PMAD) systems
were more or less the same regardless of how power was generated, with the main difference being the number of
Table 3. Point of Use Crewed Mission Power Needs.
Peak Power
Needed (W)
Keep-Alive Power
Needed (W)
Element Cargo
Phase
Crewed
Phase
Cargo
Phase
Crewed
Phase
ISRU 19,700 0 19,700 0
Mars Ascent
Vehicle 6,655 6,655 6,655 6,655
Surface Habitat 0 14,900 0 8,000
*Science
Laboratory 0 9,544 0 174
Total 26,355 31,099 26,355 14,829
*Optional element shown with all systems running. Assume
power can be phased to stay below cargo operations peak.
American Institute of Aeronautics and Astronautics
12
transmission cables accumulated over time and their associated power conversion electronics. Power Management
differences were accounted for in the mass estimates. The solar power system requires solar power and battery
regulators along with power converters and distribution electronics for the lander housekeeping and ISRU loads.
Because each solar-powered lander becomes a node in the larger network, at least one cable is required for every
lander, even after all payloads have been unloaded. In the fission power scheme, transmission cables could be
repurposed for new landers once all payloads are off-loaded from an older lander and the lander itself is
decommissioned. To estimate rough order of magnitude masses, the study team assumed 1,000 VDC transmission,
stepped down to 120 VDC at the point of use. It was assumed that transmission cables would be laid on the ground,
rather than buried.
C. Crewed Mission Concept of Operations
1. Fission Powered Concept
The concept of operations for all crewed options begins the same: launch the first of three cargo landers on an
SLS approximately two years before the first Mars expedition crew arrives on the fourth lander. In the fission-
powered scheme, all Kilopower units would arrive on the first uncrewed cargo lander. The Kilopowers could be pre-
connected together, and operate from the lander on which they
arrived, or offloaded and repositioned (Figure 9). If off-loaded,
the study team assumed the lander’s one sol, on-board, internal
energy storage was sufficient to operate a cargo off-loading
system. For the assumed shield mass, Kilopower units must be at
least 1 km from continuously occupied crew assets, such as the
habitat. Subsequent landers would have to provide internal keep-
alive power to cargo for up to one sol while power distribution
cables were robotically laid to the Kilopower farm. Some landers,
would eventually be decommissioned once all cargo was
unloaded and at that point the power distribution cable could be
repurposed for the next lander, mitigating the need to bring a new cable for each subsequent lander. The Kilopower
units would provide around-the-clock continuous power, regardless of landing site latitude, season, or dust storm
prevalence. In the event of a Kilopower failure, the spare unit would be activated. After approximately 12 years (the
equivalent of about three crewed expeditions), the Kilopower units would reach end of life and be replaced.
2. Solar Powered Concept
Upon landing, each lander would deploy up to four solar arrays to provide keep-alive power to on-board cargo.
As in the fission concept, power distribution cables would be robotically laid between landers. Every lander must
carry a distribution cable and voltage converters, because every lander is a permanent node of the power network,
even if there is no powered cargo on that particular lander, or if all powered cargo has been off-loaded. This means
that, unlike the fission system, cables and converters from a “retired” lander cannot be re-used for subsequent
landers to save mass.
In the event of a dust storm during the first expedition’s cargo phase, battery recharging would take priority
during the day, with propellant production reduced or suspended. During the crewed phase, non-essential services
could be metered to reduce power draw while operating on reduced solar collection. By the end of the second
crewed expedition, the larger solar array network could theoretically collect enough energy during a dust storm to
continue nominal operations without disruption.
D. Crewed Mission Lander Conceptual Design
1. Power Management and Distribution System
Detailed PMAD system design remains forward work, and final mass will depend on the operating voltage for
the landers and their payloads, number of landers, distance between each lander, how robust cable shielding must be,
and the transmission voltage selected. Because some equipment, such as a power distribution unit, would be part of
a lander’s flight design anyway, that function may be available to the surface power system at no additional mass
penalty. For this reason, estimated PMAD masses are addressed separately from the surface power generation
system mass.
The study team assumed 120 VDC to the main load, with a 28 VDC nominal bus for lander housekeeping loads.
A 1000 V – 120 V DC-DC converter would be carried by surface rover from one lander to the subsequent lander to
link the two together, and a 120V – 1000V DC-DC converter and power distribution unit would be located on each
subsequent lander. Table 4 envelopes the typical PMAD mass that might be expected on each solar powered lander,
Figure 9. Kilopower deployment.
American Institute of Aeronautics and Astronautics
13
given the caveats outlined above. Note that cable mass shown here is quite conservative, and could be lower
depending on insulation and abrasion protection selected. Because the Kilopower design includes a Stirling
controller/AC-DC converter, a separate 120 VDC to 1000 VDC converter would not be necessary for the fission
system’s transmission line, but a smaller converter would be required at each lander to step power down to 120
VDC for any powered payloads that remain permanently onboard the lander.
2. Crewed Mission Fission-Powered Concept
The study team found that four each 10-kWe Kilopower units would provide up to 35 kWe continuous power for
all mission phases, at either hypothetical landing site. The first lander would carry the four primary Kilopower units
plus one spare, totaling 8,769 kg (including 15% mass growth allowance). Alternating current (AC) output from
Kilopower Stirling converters would be transmitted at 1,000 VDC to either the first lander (if the Kilopower units
were relocated) or the second lander (if the Kilopower units remained on board the first lander). With 30% mass
growth allowance, Stirling cables and controllers added 385 kg to the first lander, for a total fission power
generation mass of 9,154 kg through the first crewed expedition (Table 5), not including lander-to-lander PMAD.
Thermal management and mechanisms are included in Kilopower total mass.
As noted, as much as an additional 1,038 kg PMAD mass could be needed on the first lander, if the Kilopowers
were relocated and the first lander required 1,000 VDC to 120 VDC conversion and distribution for other loads
remaining on that first lander. The second, third and fourth landers of Expedition 1 would each require a 1 km spool
of high voltage cabling, connectors, and voltage converters. Assuming cables could be retrieved and relocated, at
least two of the Expedition 1 cables could be re-used for subsequent missions (after cargo was unloaded from
Expedition 1 landers); this means that Expedition 2 and 3 landers could carry lower PMAD mass. In any case, no
additional fission power generation mass would be required for Expeditions 2 or 3. By expedition 4, the Kilopower
units would be nearing end of service life, so the first lander of Expedition 4 would carry similar mass to the first
lander of Expedition 1.
3. Jezero Crater Solar-Powered Mission Concept
The study team estimated that around-the-clock propellant production using the first two solar-powered landers
to Jezero Crater would require a maximum power load of 34,255 W during the day and 34,998 W at night, under
nominal conditions. During a dust storm, power use would be reduced to 10,985 W during the day and 11,728 W at
night. Once crew arrived, combined loads of the first four Expedition 1 landers were 31,915 W during nominal
daytime operation and 26,790 W at night; under dust storm conditions, loads dropped to 22,945 W during the day,
and 24,060 W at night.
The study team estimated that all four Expedition 1 landers would require four each 12 m diameter UltraFlex™
arrays or equivalent, with each wing estimated at 240 kg, including 30% mass growth allowance, arranged as shown
in Figure 10. Deploying arrays on a 9.1 m diameter lander would extend the overall footprint to about 33 m
diameter. With arrays in a neutral position on a 2.66 high lander deck, the overall height was about 9.69 m above the
surface. As with the demonstrator mission, arrays were deployed high to minimize interactions with the surface or
Table 4. Representative Solar Power Management and Distribution (PMAD) equipment list.
Description Quantity Unit
Mass (kg)
Basic
Mass (kg)
Growth
(%)
Growth
(kg)
Total
Mass (kg)
1 km cable + spool + connectors 1 336 336 55 184.8 521
120 VDC to 1000 VDC Converter 1 176 176 30 53 229
1000 VDC to 120 VDC Converter 1 172 172 30 51.6 224
120 V Power Distribution Unit 1 50 50 30 15 65
Total 1,038
Table 5. Jezero Crater crew Expedition 1 fission power generation mass (kg).
Description Lander 1 Lander 2 Lander 3 Lander 4
Expedition 1
Fission Power
GenerationTotal
Electrical Power Subsystem
Power Generation
10 kWe Kilopower 8,769 0 0 0
Kilopower Power Management
Stirling AC Cable 62.4 0 0 0
Stirling Controller 322.4 0 0 0
FISSION POWER SYSTEM 9,154 0 0 0 9,154 kg
American Institute of Aeronautics and Astronautics
14
payloads, and to shed dust by gimbaling (Figure 11). By using the lander deck itself as a stable operating platform,
arrays could be brought online quickly without the complication of lifting/relocating them.
As shown in Table 6 total landed solar power system generation and storage mass at Jezero Crater was estimated
at 11,713 kg over the four-lander Expedition 1, not including lander-to-lander PMAD. As with the fission power
case, as much as 1,038 kg of PMAD mass would also be required on each lander; however, unlike the fission case,
cable cannot be re-purposed without losing a node of the distributed system, so each solar powered lander would
have to carry PMAD mass, regardless of whether there are any powered payloads on board. Note that the sixteen
combined solar arrays on the four Expedition 1 landers only generate enough power to weather a dust storm in
reduced-power mode, with propellant production reduced or suspended depending on storm severity. All three
Expedition 2 landers would also have to carry solar arrays to build up enough array area to maintain normal
operation during a worst-case dust storm.
The study team estimated 5,611 kg of power generation and storage mass for the first Expedition 1 lander and
2,034 kg on each subsequent Expedition 1 lander, with the only difference being the batteries and battery
management equipment. Table 7 provides a detailed summary of power generation and distribution mass for the first
lander.
4. Columbus Crater Mission Concept
The study team found that because there are typically no dust storms in the southern hemisphere during winter,
when days are shorter, there was no need to increase solar array area from the Jezero Crater concept. However, the
longer nights at Columbus Crater did result in a 9% increase in battery mass. The study team further optimized
system performance by assuming additional ISRU strings to operate at higher capacity in summer, with some ISRU
strings cycled off all winter. This scheme would add about 40%, or 671 kg, to the ISRU system. Cumulative solar
power generation and storage mass for the four-lander Expedition 1 mission totaled 12,679 kg (Table 8), higher
than the Jezero Crater system mass. Again, lander-to-lander PMAD system mass is not included in this total.
Table 6. Jezero Crater crew Expedition 1 solar power generation and storage mass (kg).
Description Lander 1 Lander 2 Lander 3 Lander 4
Jezero Crater
Expedition 1
Solar Power
Generation and
Storage
Total
Electrical Power Subsystem 4,890 1,512 1,512 1,512
Power Generation 1,321 1,321 1,321 1,321
Lander Internal Power
Management and Distribution
401 192 192 192
Energy Storage 3,168 0 0 0
Structures and Mechanisms 660 476 476 476
Secondary Structure 416 418 418 418
Mechanisms 244 59 59 59
Thermal Control (Non-Propellant) 61 45 45 45
Active Thermal Control 2.4 3.4 3.4 3.4
Passive Thermal Control 41.8 42 42 42
Semi-Passive Thermal Control 16.8 0 0 0
SOLAR POWER SYSTEM 5,611 2,034 2,034 2,034 11,713 kg
Figure 11. Crewed mission lander, shown with
solar arrays gimbaled and a MAV payload.
Figure 10. Crewed mission lander, shown with
solar arrays deployed and a MAV payload.
American Institute of Aeronautics and Astronautics
15
Table 7. Jezero Crater crewed mission Lander 1 solar power generation and storage mass.
Description Qty. Unit
Mass (kg)
Basic Mass
(kg)
Growth
(%)
Growth
(kg)
Total
Mass (kg)
POWER SYSTEM 5,611
Electrical Power Subsystem 4,890
Power Generation
Solar Array Wing 4 200 800 30 240 1,040
Gimbal 4 54 216 30 64.8 280.8
Lander Power Management
Array Regulator 4 13 52 30 16 68
Battery Regulator 1 86 86 30 26 112
Low Voltage Power Distribution 1 10 10 30 3 13
Lander Power Harness 1 126 126 55 69 195
Low Voltage DCDC Harness 1 10 10 30 3 13
Energy Storage (Battery) 1 2,437 2,437 30 731 3,168
Structures and Mechanisms 660
Secondary Structure
Solar Array Post Assembly 4 80 319 18 57 377
Antenna Boom Assembly 2 5 10 18 2 11
Instrument Box 1 24 24 18 4 28
Mechanisms
Electrical Power Installation 1 204 204 18 37 241
Thermal Control Installation 1 2 2 18 1 3
Thermal Control (Non-Propellant) 61
Active Thermal Control
Heaters 5 0.2 1 0 0 1
Thermal Control 2 0.2 0.4 0 0 0.4
Thermal Switch 5 0.2 1 0 0 1
Passive Thermal Control
Multilayer Insulation 1 4.5 4.5 0 0 4.5
Cold Plates 5 0.1 0.5 0 0 0.5
Heat Pipes 5 1.4 6.8 0 0 6.8
Radiator 1 29.2 29.2 0 0 29.2
Thermal Paint 1 0.8 0.8 0 0 0.8
Semi-Passive Thermal Control 1 16.8 16.8 0 0 16.8
Table 8. Columbus Crater crew Expedition 1 power system mass (kg).
Description Lander 1 Lander 2 Lander 3 Lander 4
Columbus
Crater
Expedition 1
Solar Power
Generation and
Storage
Total
Electrical Power Subsystem 5,176 1,512 1,512 1,512
Power Generation 1,321 1,321 1,321 1,321
Lander Internal Power
Management and Distribution
401 192 192 192
Energy Storage 3,454 0 0 0
Structures and Mechanisms 672 476 476 476
Secondary Structure 416 416 416 416
Mechanisms 256 59 59 59
Thermal Control (Non-Propellant) 61 45 45 45
Active Thermal Control 2.4 3.4 3.4 3.4
Passive Thermal Control 42 42 42 42
Semi-Passive Thermal Control 16.8 0 0 0
POWER SYSTEM TOTAL 5,909 2,033 2,033 2,033
ADDITIONAL ISRU MASS 0 671 0 0
TOTAL 5,909 2,704 2,033 2,033 12,679 kg
American Institute of Aeronautics and Astronautics
16
5. Solar Power with Fuel Cell Backup
The study team briefly considered a third solar-powered option: pair the solar arrays with an oxygen/carbon
monoxide fuel cell instead of batteries for energy storage. In this scheme, the fuel cell would provide dust storm
keep-alive power, with solar array area sized only for nominal operations. This meant that solar arrays could be
smaller diameter, or only required on the first few landers. A duplicate of the ISRU system would produce fuel cell
reactants, which would be stored in cryogenic storage tanks during clear weather. In a dust storm, the fuel cell would
convert oxygen and carbon monoxide to energy and carbon dioxide.
Preliminary estimates assuming a 50% round trip efficiency suggested it would require four each 8.5 m diameter
solar arrays about 480 days to manufacture the 16 metric tons of reactants needed to survive just a 70-sol dust storm.
In other words, the energy required to make reactants was about the same order of magnitude as the energy required
to produce propellant, making the system less mass efficient than the other options considered.
E. Crewed Mission Solar versus Fission Comparison
1. Performance Comparison
The fission power option provides continuous power, regardless of latitude, day/night cycle, seasonal variations,
or dust storms. Given enough landers, solar power could also support continuous operation, even during a dust
storm, but will always rely on stored energy for night-time operation. Solar-powered propellant production during
the first crew expedition could be compromised by a dust storm more severe or longer than recorded observations,
resulting in loss of mission.
The fission power system would be isolated on or near the first lander, posing no interference with other surface
systems. However, because solar arrays will be mounted on all of the Expedition 1 and 2 landers, interaction with
other systems must be considered. For example, to prevent damage to the MAV during departure, arrays must be
removed or secured prior to MAV lift-off. Removing four 12 m diameter arrays from a single lander without
damaging the MAV will be challenging particularly since safe handling may only be possible at night, when the
arrays aren’t energized. To meet power needs for subsequent missions, the MAV lander arrays either need to be
relocated for continued use or replaced on a later mission, incurring additional mass penalty. If array diameter
increases for landing sites further north, it will be even more difficult to package and handle the arrays. Nominally,
crew could assist in MAV lander array relocation, mitigating the need for robotic or autonomous array handling.
However, after a six month microgravity transit to Mars, the crew will need time to recondition their muscles and
adjust to Mars gravity before being medically cleared to exert themselves in this type of Extravehicular Activity
(EVA). Mission planners must protect for the worst-case contingency where, for some reason, the surface mission
must be abandoned before crew are physically able to perform a strenuous EVA; this could potentially increase
robotic array handling mass.
2. Mass Comparison
In the fission power scheme, all the power needed to reliably produce propellant within estimated schedule
constraints and sustain subsequent crews arrives on the first lander of the first crewed expedition. Assuming a 12-
year service life for the Kilopower units, this power generation mass will support a three-expedition campaign.
Recall that the sixteen combined solar arrays on the four Expedition 1 landers only generate enough power to
weather a dust storm in reduced-power mode, with propellant production reduced or suspended depending on storm
severity. Therefore, comparing solar vs. fission power generation mass for only Expedition 1 would be misleading.
At Jezero Crater, all three Expedition 2 landers would also have to carry solar arrays to build up enough area to
match fission power’s ability to operate at nominal levels during a worst-case dust storm, plus 10 kW of spare
power. Therefore, extrapolating cumulative power generation mass over the first three crewed expedition missions
provides a more realistic comparison of the two schemes.
Unlike the equatorial demonstrator mission, solar power does not trade as well for mass at higher latitudes.
Assuming the first two crewed expeditions to Jezero Crater accumulate sufficient solar power for nominal
operations even in a dust storm, landed solar power generation and storage mass is nearly double that of a
comparable fission power system; the discrepancy only increases as the landing site moves further from the equator
(Table 9).
A fourth crewed expedition would incur additional mass penalty to replace aging Kilopower units, but the solar
power system’s three metric tons of rechargeable batteries (the mass equivalent of two Kilopower units) may also
need to be replaced at that point.
3. System Robustness
American Institute of Aeronautics and Astronautics
17
If the campaign is limited to a single landing site,
power system mobility will be key to exploring
further from the home base. As with the
demonstrator mission, dust is a major concern for
solar array mechanisms. The compact Kilopower
units may have an operational advantage because
they can be easily relocated with minimal
reconfiguration whereas it may be difficult to re-stow
a large, dusty solar array for surface transport and
redeployment. Solar arrays are generally not
designed to be retractable in a dusty environment.
Large solar arrays will also be more sensitive to
Martian wind loads than the relatively compact, low
center of mass Kilopower units. On the other hand,
solar power system electronics will not be exposed to
system-generated radiation, and solar arrays can be
relocated as soon as they are deactivated; a 10-kWe
Kilopower unit won’t be safe for crew to approach
until about a week after it’s turned off, though it
could be robotically handled within one sol.
One distinct advantage of solar power is that if
every lander were fitted with arrays, each lander
could be independent for keep-alive power shortly
after landing, minimizing the time pressure to
robotically attach surface power system cables to
each new lander as it arrives. With up to four arrays
per lander, a dust-induced gimbal failure or
degradation of a single wing could be easily tolerated
with array area accumulated by Expedition 3. Solar
power generation closer to the point of use also
reduces transmission losses, mitigating the need to generate more power than what is actually needed.
If fission-powered landers are daisy-chained together via a single transmission cable back to a Kilopower farm,
there is a risk of disruption due to a severed transmission cable. Risk can be mitigated by adding a second cable, but
at a mass penalty. Solar arrays on every lander would make the system less vulnerable to a severed power
distribution cable—at least during the day—but the downside is that each lander has to give up cargo space for array
mass. Loss of the transmission cable back to the lander that houses the night-time energy storage batteries poses a
significant night-time risk, though this could be mitigated by distributing battery mass among landers.
4. Cost Comparison
A rough order of magnitude cost estimate was performed for the crewed mission solar power generation and
storage system but not for the fission system. Kilopower fission system development costs have been estimated in
previous publications16. As shown in Table 10, the solar power generation and storage hardware for the first lander
of the first crewed expedition is estimated at about $456 M. Because the first lander carries all of the energy storage
mass, subsequent landers were estimated to have lower costs: $196 M for Lander 2, $188 M for Lander 3, and $194
M for crewed Lander 4. Costs for power systems on Expedition 2 landers 1, 2, and 3 would be the same as, or
slightly lower than, Expedition 1 landers 2, 3, and 4. Solar power cost estimates include only the power system and
directly related structures and thermal control for the first lander. PMAD costs were removed for simplified
comparison. Launch services, mission-level costs, other payloads (such as ISRU), payload and system integration,
and technology development up to TRL 6 were not included. Estimate assumed prototype development for the solar
arrays, gimbals, and batteries and protoflight development of remaining components. Flight spares include solar
arrays, gimbals, batteries, power electronics cards, and power system thermal control components. No flight spares
or test hardware were included in the estimates for power generation on landers 2-4. Quantitative risk analysis was
performed using a Monte Carlo simulation driven by input parameter uncertainty and the error statistics of the cost
estimating relationships. Costs presented are mode values (approximately 35th percentile) in constant FY16 U.S.
dollars. The omitted PMAD costs (DDT&E and FHW) are estimated to be an additional $124.5M.
Table 9. Mass comparison for 3-mission crewed
campaign power generation and storage options (not
including lander-to-lander power distribution mass).
Crew
Expedition
Power Generation/Storage Mass (kg)
Fission
Power
Solar Power
Jezero
Crater
Columbus
Crater
Expedition 1 9,154 11,713 12,679
Lander 1 9,154 5,611 5,909
Lander 2 0 2,034 *2,704
Lander 3 0 2,034 2,033
Lander 4 0 2,034 2,033
Expedition 2 0 6,102 6,770
Lander 1 0 2,034 *2,704
Lander 2 0 2,034 2,033
Lander 3 0 2,034 2,033
Expedition 3 0 0 0
Lander 1 0 0 0
Lander 2 0 0 0
Lander 3 0 0 0
Three Mission
Total (kg) 9,154 17,815 19,449
*Columbus Crater totals include additional ISRU strings on
MAV landers.
American Institute of Aeronautics and Astronautics
18
Although Fission power system development and flight certification costs, given the required development,
launch safety and security overhead, are higher than solar power costs, there will still be a substantial cost to design,
develop, test, and flight certify a crew-rated solar power system compatible with the Mars surface environment.
5. Operational Life Comparison
There is insufficient data for a comprehensive operational life comparison. A 12 year service life is assumed for
the Kilopower units, but no life test data is currently available. With few moving parts, mechanical failures are less
of a concern than system electronics operational life.
NASA’s Opportunity rover has demonstrated solar panel and rechargeable battery performance over a 12 year
service life on Mars—but its twin, Spirit, survived only half that long17. Lithium sulfur batteries were selected for
this study due to their high specific energy and projected future improvements, though they currently have relatively
poor service life18. To put a Mars mission in perspective: a three-expedition Mars service life equates to about 8,244
battery day/night charge/discharge cycles, but Lithium batteries are currently capable of only about 6,000
charge/discharge cycles at 100% depth of discharge19. At a more reasonable 75% depth of discharge, these batteries
may last 10,000 cycles20.
V. Discussion
1. ISRU Demonstrator Mission
By assuming an equatorial landing site, ISRU demonstrator mission analysis was purposely skewed towards
solar power as more mass efficient; note that the trade is expected to quickly swing towards fission power at sites
further from the equator. This analysis was also skewed against fission power, since the study team assumed a
baseline 10-kWe Kilopower unit, which was about 20% over-sized for the 1/5 scale demonstrator. What’s more,
about a third of a nominal Kilopower unit’s mass is dedicated to crew radiation shielding, which is not strictly
necessary for an uncrewed demonstration. Since both the fission and solar concepts fell comfortably within
allowable mass limits for the demonstrator lander, mass alone is unlikely to be a delineator for this type of mission.
A final decision will hinge on many factors beyond the scope of this analysis, including technology investment
strategies, program budgets, and risk mitigation needs for later crewed missions.
2. Crewed Mission
50 kWe of fission power generation (four primary 10-kWe units plus a fifth spare) is about 20% less landed mass
than 35 kW of solar power generation and storage for the first expedition to Jezero Crater, not including a lander-to-
lander PMAD system; adding PMAD further increases fission power’s mass advantage over solar power,
particularly if Kilopower transmission cables can be re-used as previous landers are decommissioned. By the third
crewed expedition mission to Jezero Crater, enough solar array area will have been accumulated to match the fission
system’s performance, even in a dust storm--but at more than twice the cumulative landed mass of the fission
system, not including PMAD. Landing closer to the equator would make solar more competitive with fission for
landed mass, while moving further from the equator would make solar power less competitive.
Table 10. Lander 1 of Expedition 1: solar power generation and storage estimated costs (FY16$M)
Description
Design &
Development
Estimate
System
Test
Hardware
Flight
Spares
Design,
Development, Test
& Evaluation Total
Flight
Hardware
Total
Total
Electrical Power
Subsystem 105.5 42.8 45.4
193.8 224.8 418.6
Solar Array Wing 72.8 33.7 33.7 140.3 135.0 275.3
Gimbal 6.2 6.4 6.4 19.0 25.8 44.8
Lander Power Harness 8.2 0.0 0.0 8.2 5.6 13.8
Battery 18.3 2.7 5.3 26.3 58.4 84.7
Thermal Control (non-
Propellant) 4.1 0.0 0.3
4.3 2.6 6.9
Structures and
Mechanisms 10.4 0.0 0.0
10.4 19.8 30.3
System Subtotal 120.0 42.8 45.7 208.5 247.2 455.8
Does not include lander-to-lander power management and distribution hardware costs.
American Institute of Aeronautics and Astronautics
19
The two landing site point solutions that were studied offer some insight into solar power mass sensitivity with
latitude, but it remains forward work to evaluate the full range of sites under consideration. From a resource
prospecting point of view, sites further from the equator appear more likely to feature water ice21—but solar power
mass penalty also increases as a function of distance from the equator. Northern versus southern hemisphere
seasonal variations must also be considered. The Columbus Crater analysis demonstrates that overall mission mass
can be minimized with a careful systems engineering approach that takes advantage of operational schemes to offset
environmental constraints.
Because lander-to-lander power transmission mass was assumed to be comparable between the two power
generation systems, it remains forward work to develop a detailed PMAD design that optimizes transmission voltage
and meets end user needs. At an estimated metric ton per lander, PMAD mass represents a serious potential mass
threat.
Note that masses shown are for stand-alone systems, but may change once more detailed lander integration
analysis is performed. Thermal and structural support mass for the solar arrays may be stripped out if the lander
already carries mass to support these functions (lander internal cold plate function shifts from in-flight systems to
power generation after landing, for example). Also note that analysis was based on a single point design solution;
many factors could alter the analysis. If propellant production rates increase, either because more propellant is
needed or there is less time to make it, power generation and energy storage mass will increase.
As with the demonstrator mission, a final decision regarding Mars surface power generation will depend on
many factors beyond the scope of this analysis, including program budgets, technology investment strategies, and
crew or mission risk posture.
VI. Conclusions
ISRU demonstrator mission analysis indicates that solar power is more mass-efficient than fission power, at least
for an equatorial landing site and a somewhat oversized 10-kWe Kilopower unit. A solar-powered concept
operating only during daylight hours offers the lowest mass at 1,128 kg, but takes twice as long to meet fission’s
performance, and potentially has a higher risk of ISRU reliability issues due to daily on/off cycling. An around-the-
clock solar-powered system was estimated at 2,425 kg, or about 326 kg less than a fission-powered system of
comparable performance. All options evaluated fit within the mass envelope of a commercial launch platform,
making mass less of a delineator between them. The fission-powered system’s flight hardware costs beyond TRL 6
were estimated at $428M, about $100M more than the comparable solar-powered system. Final selection will
depend on long-term technology investment strategies and priorities: is it more important to demonstrate
technologies needed for a future crewed mission, or to keep mass as low as possible?
Crewed mission analysis indicates that solar power is more feasible under NASA’s new Evolvable Mars
Campaign (where surface assets are re-used for subsequent expeditions) than it was under the DRA 5.0 scheme. The
mission concept developed by the COMPASS team demonstrates that a solar-powered crew mission is certainly
possible, though not as mass-efficient as fission power for the two landing sites analyzed. The study team’s
approach of adding solar arrays to multiple landers helps spread the mass penalty out over time, but requires
program managers to accept a higher loss-of-mission risk for at least the first crew expedition (and possibly more,
depending on latitude selected and propellant production timelines). As solar array area is accumulated, mission
risk is substantially reduced. This approach makes higher latitude sites more accessible for solar power, though it
remains forward work to evaluate cumulative array area needed for all landing sites of interest. An additional benefit
of lander-mounted solar power is that it makes that lander self-sufficient for keep-alive power shortly after landing,
without the time pressure of autonomously connecting cables to a remote power farm before on-board energy
storage is depleted.
Cumulative fission power generation mass over the first three crewed expeditions to Mars is estimated at 9,154
kg, compared to 17,815 kg for a comparable solar power generation and storage system at Jezero Crater (not
including lander-to-lander distribution and conditioning mass); 19,449 kg would be required for a Columbus Crater
solar-powered mission. Clearly, the method of surface power generation will be an important factor in landing site
selection (or vice versa). A comprehensive cost comparison remains forward work, as it depends on many factors
such as landing site, surface element design and power needs, potential cost-sharing between flight programs, and
technology investment strategies. Development of either solar or fission surface power for Mars would require a
substantial financial commitment.
Pairing solar arrays with fuel cells seems logical, particularly if in situ resources are available, but the study team
concluded that the energy required to make reactants was about the same order of magnitude as the energy required
American Institute of Aeronautics and Astronautics
20
to produce propellant, making the system less mass efficient than the other options considered. If rechargeable
battery service life requires frequent replacement, fuel cells may trade more favorably than solar power over a multi-
expedition timeframe.
Acknowledgments
The authors thank Dr. Roger Meyer for suggesting we perform this study for the sake of due diligence, and Dr.
Steve Hoffman and Kevin Watts for invaluable technical support.
References 1Drake, B. G. (ed.), “Human Exploration of Mars, Design Reference Architecture (DRA) 5.0,” NASA-SP-2009-566,
National Aeronautics and Space Administration, July, 2009. 2NASA/TM—2010-216772, “Fission Surface Power System Initial Concept Definition,” Fission Surface Power Team,
National Aeronautics and Space Administration and Department of Energy, Washington, August, 2010. 3Rucker, M. A., “Integrated Surface Power Strategy for Mars,” Nuclear and Emerging Technologies for Space (NETS) 2015,
Albuquerque, 2015. 4Mason, L., M. Gibson, and D. Poston, NASA/TM—2013-216541, “Kilowatt-Class Fission Power Systems for Science and
Human Precursor Missions,” National Aeronautics and Space Administration, September, 2013. 5Craig, D.A., P. Troutman, N.B. Hermann, AIAA 2015-4409, “Pioneering Space Through an Evolvable Mars Campaign,”
AIAA Space 2015 Conference and Exposition, Pasadena, 2015. 6Yount, B., D. Prabhu, C. Kruger, P. Gage, T. Squire, and J.A. Mockelman, “Structures and Mechanisms Design Concepts for
Adaptive Deployable Entry Placement Technology,” AIAA 2013-1369, AIAA Aerodynamic Decelerator Systems (ADS)
Conference, Daytona, March, 2013. 7Wang, H. and M.I. Richardson, “The origin, evolution, and trajectory of large dust storms on Mars during Mars years 24–30
(1999–2011),” Icarus International Journal of Solar System Studies, Vol. 251, pp. 112-127, 2013. 8Lemmon, M. T., et al., “Dust aerosol, clouds, and the atmospheric optical depth record over 5 Mars years of the Mars
Exploration Rover mission," Icarus, Vol 251, pg 96-111, May 1, 2015. 9NASA Press Release 2007-080 (July 20, 2007), “NASA Mars Rovers Braving Severe Dust Storms,” National Aeronautics and
Space Administration, http://mars.nasa.gov/mer/newsroom/pressreleases/20070720a.html [cited August 14, 2016]. 10Sanders, G. B., Paz, A., Oryschchyn, L., Araghi, K., Muscatello, A. C., Linne, D. L., Kleinhenz, J. E., and Peters, T., “Mars
ISRU for Production of Mission Critical Consumables—Options, Recent Studies, and Current State of the Art,” AIAA Space
2015, AIAA 2015-4458, American Institute of Aeronautics and Astronautics, Pasadena, 2015. 11McNatt, J.S., G.A. Landis, and J. Fincannon, “Design of Photovoltaic Power System for a Precursor Mission for Human
Exploration of Mars,” 43rd IEEE Photovoltaic Specialists Conference, Portland, June 3-8, 2016. 12AIAA Standard, “Standard Mass Properties Control for Space Systems,” S–120–2006, American Institute of Aeronautics
and Astronautics, 2006. 13“NASA Space Flight Program and Project Management Requirements,” NPR 7120.5E, National Aeronautics and Space
Administration, Washington, 2012. 14“Prospective Landing Sites for Future Missions,” National Aeronautics and Space Administration,
http://marsoweb.nas.nasa.gov/landingsites/ [cited May 2, 2016]. 15Mars 2020 Landing Site Workshop, National Aeronautics and Space Administration,
http://marsnext.jpl.nasa.gov/workshops/wkshp_2015_08.cfm [cited May 2, 2016]. 16”Nuclear Power Assessment Mission Studies for Enabling and Extending Future Space Exploration Final Report,” RPS-
RPT-0121 (JPL D-81712), National Aeronautics and Space Administration, 2015. 17Update: Spirit and Opportunity, Mars Exploration Rovers, NASA Jet Propulsion Laboratory,
http://mars.nasa.gov/mer/mission/status.html [cited August 9, 2016]. 18Song, M.K., Y. Zhang, and E.J. Cairns, “A Long-Life, High-Rate Lithium/Sulfur Cell: A Multifaceted Approach to
Enhancing Cell Performance,” Nano Lett., 13 (12), pp. 5891–5899, 2013, 19Gutsch, A., “PV-Speichertests durch das KIT,” 30th Photovoltaische Solarenergie Symposium, Kloster Bans, Bad
Staffelstein, March, 2015. 20Doring, H., “Performance und Sicherheit von PV-Li-Speicherbatterien”, 31st Photovoltaische Solarenergie Symposium,
Kloster Banz, Bad Staffelstein, March, 2016. 21Christensen, P. R., "Water at the Poles and in Permafrost Regions of Mars," GeoScienceWorld Elements, 3 (2) pp. 151–
155, 2006.